NoC-Armor: Leveraging Quantitative Analysis for Enhanced Security

In the realm of multi-processing systems, security challenges persist with the increasing interconnectivity of devices, especially within Network-on-Chip (NoC) architectures. Shared hardware resources between secure and malicious IPs create vulnerabilities exploited by various attacks. Side-channel attacks pose challenges due to their exploitation of implicit information leakages, relying on power consumption, electromagnetic radiation, or timing variations. In response, we introduce NoC-Armor that dynamically identifies and counters side-channel attack patterns within NoCs. NoC-Armor employs a quantitative approach, utilizing conditional probability as a metric, to examine secure and side-channel attack events. This method serves as a valuable tool for making routing decisions, ultimately enhancing the security posture of NoC architectures. Our methodology demonstrated a 50% enhancement in security resilience with minimum power and area overhead.